Top Navigation Bar

ASAE Journal Article

Removal Efficiency of Horizontal Subsurface Flow Wetlands for Veterinary Pharmaceuticals

S. A. Hussain, S. O. Prasher, R. Patel


Published in Transactions of the ASABE Vol. 54(6): 2037-2046 ( Copyright 2011 American Society of Agricultural and Biological Engineers ).

Submitted for review in May 2011 as manuscript number SW 9213; approved for publication by the Soil & Water Division of ASABE in September 2011.

The authors are Syed Azfar Hussain, Graduate Student, Shiv O. Prasher, CSABE Fellow, Professor and Chair, and Ramanbhai Patel, Research Associate, Department of Bioresource Engineering, McGill University, Ste. Anne de Bellevue, Quebec, Canada. Corresponding author: Shiv O. Prasher, Department of Bioresource Engineering, McGill University, 21111 Lakeshore Road, Ste. Anne de Bellevue, Quebec, Canada H9X 3V9; phone: 514-398-7775; fax: 514-398-8387; e-mail: shiv.prasher@mcgill.ca.


Abstract. This study was undertaken to assess the removal of commonly used antibiotics on poultry and livestock farms, i.e., monensin, salinomycin, and narasin, in a horizontal subsurface flow (HSSF) treatment wetland. The tested wetland supported a stand of reed canary grass ( Phalaris arundinaceae L.) and cattails ( Typha latifolia L.) on sandy soil. Of the three compounds evaluated, monensin was found to be the most mobile, and the wetland exhibited an average removal efficiency of 40%. The removal rates for both salinomycin and narasin approached 50%. The effect of the applied concentration of antibiotics on wetland removal efficiency was found to be highly significant (p < 0.0001) only at the first two concentration levels (10 and 50 µ g L -1 ) versus the two higher levels (0.5 and 1.0 mg L -1 ). Even though raw data on temperature, dissolved oxygen (DO), oxidation reduction potential (ORP), and pH suggested a microbial contribution to overall antibiotics removal, the effect of these parameters was found to be non-significant (p < 0.05). However, the removal trends were found to be in line with our previously reported laboratory work on biodegradation of these pharmaceuticals.

Keywords . Antibiotics, Horizontal subsurface wetlands, Ionophores, Pharmaceuticals, Removal efficiency.

Antibiotics are used in the livestock industry to achieve therapeutic, prophylactic, and growth promotion objectives. The Animal Health Institute reported that 12,000 tons of antibiotics were sold for animal use in the U.S. in 2006 (AHI, 2007). Among the pharmaceuticals used exclusively for veterinary purposes, ionophores are a prominent class. These compounds are comprised of complex, high molecular weight molecules derived from various Streptomyces species. Among this group, monensin, salinomycin, and narasin are some of the most commonly administered drugs. High usage of these drugs has resulted in their corresponding release to the environment. A number of recent studies have detected these xenobiotics in different environmental matrices (Onesios and Bouwer, 2009; Watanabe et al., 2008). Despite the fact that some studies have found these compounds to be somewhat susceptible to microbial degradation (Ramaswamy et al., 2010; Hussain et al., 2011), based on their usage and detected environmental concentrations, these antibiotics are considered to be high-risk compounds (Hansen et al., 2009). Moreover, because of their antimicrobial characteristics, these antibiotics can also kill pollutant-specific degrading bacteria, thus increasing pollutants' half-life (Kim et al., 2011).

In the context of agricultural pollution control, constructed wetlands (CWs) have been found to be a viable treatment option that not only blends in well with rural environments but also has the ability to effectively handle the highly unpredictable and variable loading rates of agricultural effluents. CWs can arguably perform with little loss of efficiency under the variable volumes of water and varying contaminant levels that are typical of agricultural runoffs (Sim et al., 2008). This passive treatment approach has been shown to be highly effective in removing various forms of organic and inorganic pollutants, while being substantially economical with minimum maintenance and running cost (Kadlec and Wallace, 2008). Complex ecosystems of plants, microorganisms, and substrate in CWs act together as a biogeochemical filter that can efficiently remove low levels of contaminants from large volumes of water (Kosolapov et al., 2004).

Although a number of research studies have been carried out on pollutant removal by wetland ecosystems, use of constructed wetlands for these emerging contaminants is relatively new, with only a few studies undertaken in this field. These studies appraised the efficiency of CWs for various types of pharmaceuticals and personal care products (PPCPs) like endocrine disruptors (Chapman, 2003; Fisher and Scott, 2008), hormones (Gray and Sedlak, 2005), human antimicrobials (Conkle et al., 2008; Ávila et al., 2010; Conkle et al., 2010), and veterinary antibiotics (Kuchta et al., 2009; Matamoros et al., 2008a; Matamoros et al., 2008b). The removal of 30 different micropollutants, including pharmaceuticals, endocrine disruptors, and personal care products, was evaluated in an engineered constructed wetland, and only nine of the 30 chemicals were found to be resistant to removal (Park et al., 2009). A recent study compared the removal efficiency of three different full-scale hybrid pond-constructed wetlands and a conventional wastewater treatment plant (WWTP) for PPCPs present in primary-treated urban wastewaters. The hybrid systems were found to be at least as efficient as the WWTP, with removal efficiencies exceeding 70% (Hijosa-Valsero et al., 2010a).

On the basis of flow regimes, treatment wetlands can be broadly divided into three major categories: free water surface (FWS), horizontal subsurface flow (HSSF), and vertical flow (VF). FWS and HSSF are the most commonly employed wetland types. Wetland type selection depends on regional practices, climatic considerations, the nature of the pollutants and their loads, aesthetics, health and wildlife concerns, land and topographical limitations, etc. While some studies have reported comparatively better treatment efficiencies for HSSF systems (Van de Moortel et al., 2009), others have found that HSSF and FWS wetlands may not be significantly different in their treatment performance (Kadlec and Wallace, 2008). Matamoros et al. (2007) indicated that for the removal of a number of PPCPs, subsurface vertical flow CWs exhibited higher treatment efficiencies than CWs with other configurations (horizontal subsurface flow and surface flow).

Therefore, the primary objective of this study was to evaluate and quantify the removal efficiency of an HSSF system for a range of concentrations of selected pharmaceuticals, expected in agricultural effluents, emanating from recently manured fields or confined animal farming operations (CAFOs). The three selected antibiotics are widely used for both poultry and livestock, and until now no comprehensive constructed wetland study has been undertaken to evaluate their removal potential in such a treatment setup. Moreover, this work could also throw some light on the preferred microbial degradation pathway for the tested compounds. Although a number of studies have indicated aerobic degradation as the preferred pathway for removal of certain pharmaceuticals (Matamoros et al., 2008a; Hijosa-Valsero et al., 2010b; Musson et al., 2010), others have found anaerobic pathways to be more efficient (Ávila et al., 2010).

Materials and Methods

Wetland System and Field Setup

The wastewater treatment study site was located 3 km north of McGill University's Macdonald Campus in Ste. Anne de Bellevue, Quebec, Canada. The CWs were built in triplicate using three above-ground, open half-cylinder tanks made of black high-density polyethylene (HDPE). The tanks were filled with a sandy soil. The relevant characteristics of the sandy soil used in the study are given in table 1. Prior to the current experiment, the tanks were used as FWS wetlands. For a more comprehensive investigation of the removal potential of CWs, the study was broadened to include an HSSF system, too. In each of the CWs, an outlet was made at a depth of 20 cm from the soil surface to create an HSSF wetland system (fig. 1). A filter made of two mesh sizes (1.905 and 0.635 cm) of clean and washed gravel was placed at the outlet to prevent clogging.

SW9213_files/image1.jpg

Figure 1. (a) Dimensions of the wetland; (b) and (c) are top and side sectional views of the constructed wetland (adapted from Yates and Prasher, 2009).

Table 1. Characteristics of sandy soil used in the wetlands.

Sand (%)

99

Silt (%)

0.2

Clay (%)

0.8

Organic matter (%)

4.7

Al (mg kg -1 ) [a]

335.3 (9.5) [b]

Fe (mg kg -1 ) [a]

57.6 (5.1) [b]

[a] Al and Fe determined by ICP (AES).

[b] Standard deviations are shown in parentheses.

Each CW tank had a volume of approximately 4.17 m 3 and surface area of 9.29 m 2 . At the inlet end of the CW, a transverse deep zone was provided to stabilize the incoming flow, avoid short-circuiting, induce lateral mixing, and help in attaining a uniform distribution of water in the soil profile (fig. 2). A recent modeling study suggested that deep zones, when properly sized and located, may enhance wetland performance (Lightbody et al., 2009). Approximately 66% of the surface area of the wetland contained soil and was vegetated with alternate bands of reed canary grass ( Phalaris arundinaceae L.) and cattails ( Typha latifolia L.) (fig. 1). Some of the cattails had to be transplanted each year from a neighboring pond to fill gaps in the wetlands.

SW9213_files/image2.jpg

Figure 2. The constructed wetlands.

The field setup for the experiment is shown in figure 3. A continuous-flow delivery system was established to supply water prepared with the desired pollutant concentrations to the three replicated CWs at a daily mean flow rate of 1.0 L min -1 . However, the actual average supply was 0.8 L min -1 . The mean hydraulic residence time for the CWs, as ascertained through a bromide tracer study, was 4.06 days.

The prepared mixture was stocked in three reservoir tanks, each having a capacity of 9,000 L, and delivered to the CWs via distribution tanks. The water was supplied to the CWs through an inflow manifold. Inflow distribution manifolds have been shown to improve hydraulic retention time (Shilton and Prasad, 1996). Constant flow was maintained by using multiple flow-control valves. Actual removal of the three antibiotics, for each concentration level, was determined by calculating the difference between the inflow and outflow concentrations of the ionophoric antibiotics.

SW9213_files/image3.gif

Figure 3. Schematic diagram of the treatment wetland system (not to scale).

A tipping bucket was installed at the outlet of each experimental unit to record actual flow (fig. 1). All tipping buckets were connected to dataloggers. The effluent from the treatments was drained into a sump from where the water was pumped out to be spread over a large uncultivated field.

The study was conducted for a period of two years from late June to early October (2008-2009). The pre-study sampling did not find detectable concentration of the evaluated antibiotics. Average soil temperatures for the three months of study were 19.2°C, 19.7°C, and 12.9°C for the first year and 19.1°C, 20.7°C, and 13.0°C for the second year. Dissolved oxygen (DO), pH, oxidation-reduction potential (ORP), and temperature data were collected biweekly in the deep zone and near the outlet. These data were collected using portable pH meter (model D-52, Horiba, Ltd., Kyoto, Japan) with a corresponding DO probe, ORP probe, and pH/temperature probe. All probes were calibrated and stored according to Horiba's specifications.

Chemicals

Analytical standards of monensin, salinomycin, and narasin were bought from Sigma-Aldrich (St. Louis, Mo.) in the form of sodium salts with 95% to 98% purity. The relevant characteristics of these compounds are given in table 2. Mobile-phase chemicals were procured as follows: HPLC-grade methanol was obtained from EMD Chemicals (Gibbstown, N.J.), and acetic acid (glacial) and ammonium hydroxide were obtained from Fisher Scientific (Fair Lawn, N.J.). Double-deionized water (Milli-Q, Millipore, Molsheim, France) was used in the study. Nigericin was used as an internal standard and was obtained from Sigma-Aldrich (St. Louis, Mo.). Stock solutions of 100 mg L -1 of the three ionophoric compounds (monensin, salinomycin, and narasin) were made in HPLC-grade methanol, whereas the working standards were prepared in HPLC-grade methanol on a bi-weekly basis at concentrations of 0.01, 0.1, 1.0, 5.0, and 10 mg L -1 . The stock solutions and the working standards were both stored under dark in amber-colored glass bottles at 4°C. The buffer for the LC/MS mobile phase was prepared by mixing 0.05 M glacial acetic acid with 0.05 M ammonium hydroxide to attain a pH of 5.0. The solution was autoclaved and stored for a maximum period of one week. Bromide for the tracer study was procured from Fisher Scientific (Fair Lawn, N.J.).

Table 2. Physical and chemical characteristics of monensin, salinomycin and narasin antibiotics.

Compound

Monensin

Salinomycin

Narasin

CAS No.

17090-79-8

53003-10-4

55134-13-9

MW (g mol -1 )

671

751

765

Formula

C 36 H 62 O 11

C 42 H 70 O 11

C 43 H 72 O 11

Structure [a]

SW9213_files/image4.jpg

SW9213_files/image5.jpg

SW9213_files/image6.jpg

Produced by

Streptomyces cinnamonensis

Streptomyces albus

Streptomyces auriofaciens

Solubility in water

4.8 to 8.9 [b] , <100 [k] mg L -1

57 to 905 [d] , 302 to 7685 [c] mg L -1

102 to 681 [g] mg L -1

pK a

4.2 [e] or 6.65 [b]

4.4 [e],[f]

4.4 [e] or 7.9 [g]

Stability

Unstable in acidic condition, stable in alkaline condition

Unstable in acidic condition, stable in alkaline condition

Unstable in acidic condition, stable in alkaline condition

log K oc

2.1 to 3.8 [h] or >5.6 [b]

2.2 to 2.8 [d]

6.06 to 6.88 [g] , >5.63 [j]

log K ow

2.8 to 4.2 [b] , >6.3 [b] , or 5.4 to 8.5 [l]

5.15 [d] , >6.2 [c]

4.85 to >6.2 [g]

Soil DT 50

3.8 [h] , 7.5 [m] , or 13 to 18 days [b]

5 [i] or 8 to 18 days [d]

8.8, 21 to 49 days [j]

[a] Kim and Carlson (2006).

[b] EFSA (2004a).

[c] EFSA(2004b)

[d] EFSA (2004c).

[e] Calculated using Marvinsketch (www.chemaxon.com/marvin/index.html).

[f] Carmosini and Lee (2008).

[g] EFSA (2007).

[h] Sassman and Lee (2007).

[i] Schlüsener and Bester (2006).

[j] Elanco (2004).

[k] Dolliver and Gupta (2008a).

[l] Thiele-Bruhn (2003).

[m] Carlson and Mabury (2006).

Bromide Tracer Study

A bromide tracer study was conducted on the HSSF wetlands to determine the mean retention time. Bromide was chosen because of its inert properties and high recovery rate, as demonstrated in other wetland tracer studies (Yates and Prasher, 2009). As an added protection, the experiment was performed when the vegetative growing season was over (senescence and abscission of plants had started) and the temperature had sufficiently cooled to minimize any chances of plant uptake and biological dissipation of the tracer used. The mass of bromide added was 1,000 times the detection limit of 0.1 mg L -1 , multiplied by the wetland volume, and was therefore 417 g for each CW. Once steady-flow conditions were achieved, the bromide (ACS-grade potassium bromide) was dissolved in 4 L of distilled water and was added to the inflow of each CW. Outflow water samples were collected every 3 h for each wetland until the concentration of bromide became undetectable or reached the background concentration level. Outflow data were recorded throughout the study period using tipping buckets connected to dataloggers. Samples were analyzed in the laboratory for bromide concentration using the bromide colorimetric method (APHA, 1992).

Mean residence time was calculated using chemical reactor residence time distribution theory, following the procedure of Yates and Prasher (2009). In order to compensate for the variable flow experienced in the CWs, the bromide concentration detected for each sampling time was multiplied by the corresponding flow, as shown in the following equation:

C = C 0 * Q (1)

where C is the bromide concentration (g per 3 h), C 0 is the instantaneous sample bromide concentration (g L -1 ), and Q is the inflow/outflow rate (L per 3 h). Once the concentration was converted into mass per unit time, it was possible to integrate the area under the curve to find the total amount of tracer recovered. Observations taken for a period of 10 days were used to calculate the retention time:

SW9213_files\eqn\eqn1.gif (2)

where A is the area under the curve, i is the sampling number (1, 2, 3, 4, ..., 80), ? t is the time elapsed between sampling (3 h), and C i is the grams of bromide from the i th 3 h time unit. The mean residence time (RT) was then calculated by:

SW9213_files\eqn\eqn2.gif (3)

where t is the unit time (3 h).

Influent Preparation

To prepare the solution of desired concentration for field application, feed additive premixes of salinomycin, monensin, and narasin containing 60, 200, and 70 g kg -1 of active ingredient, respectively, were used. Pre-weighed amounts of these supplements were mixed with ACS-grade methanol procured from Fisher Scientific (Fair Lawn, N.J.) in a 1:5 ratio and left for 4 h on a rotary shaker at 150 rpm. The mixture was then subjected to repeated washing with ACS-grade methanol to extract the active ingredients. A recovery of 60% to 65% was achieved by this method. Concentration levels for the pharmaceuticals were selected keeping in consideration the high concentration discharges possible from CAFOs and freshly manured farmer's fields. Dolliver and Gupta (2008a) reported a maximum concentration of 3175 µ g L -1 for monensin in runoff from manure stockpiles that are typically present near CAFOs. A second study reported monensin losses of 40.9 and 57.5 µ g L - 1 in leachate and runoff, respectively (Dolliver and Gupta, 2008b). In another study, while evaluating field losses of different antibiotics, monensin was found to have the highest concentration in runoff water (Davis et al., 2006). With an aim to cover the possible concentration range in agricultural effluents, five levels of antibiotics concentrations using tap water were employed for this study: 10, 50, 100, 500, and 1000 µ g L -1 . Once the desired concentration was formulated in the field tanks, the water was left standing for at least 24 h before being supplied to the CWs. Each concentration was supplied to the CWs for a period of one month. Based on the expected retention time of the wetlands, an equilibration time of seven days was given between each level.

Sampling

Sample collection was handled with a custom-built Avensys auto-sampler with a built-in sample storage refrigerator maintained at 4°C. The auto-sampler was provided with an auto-channel switcher to switch from one treatment to the next, once the sample for the specific time had been collected. The sampler was programmed to take samples at 42 h intervals. The samples were combined to make a composite sample on weekly basis that represented the corresponding period of effluent flow. The composite samples were collected in 1 L amber-colored glass bottles, and a pre-specified amount of internal standard nigericin was added (Hussain and Prasher, 2011). Grab samples were also collected at each sampling point and date to verify the validity of the sampling procedure.

Filtration and Solid Phase Extraction

After collection, samples were subjected to a two-step filtration process, first with a 1.2 µ m glass fiber filter (Fisher Scientific, Fair Lawn, N.J.) and then with a 0.45 µ m glass fiber filter (Fisher Scientific, Fair Lawn, N.J.). As a final cleaning and concentration step, solid phase extraction was performed using 60 mg/3.0 mL Oasis HLB cartridges on a 20-slot Waters vacuum manifold (Waters, Ltd., Lachine, Quebec, Canada). The cartridges were conditioned first with 5.0 mL of HPLC-grade methanol, next with 5.0 mL of 50:50 methanol:water, then with 5.0 mL water at pH 4 (using hydrochloric acid), and lastly with 5.0 mL of double-deionized Milli-Q water.

A flow rate of 5.0 mL min -1 was maintained for the solid phase extraction step. After loading, the cartridge was air-dried and ionophores were eluted with 5.0 mL of HPLC-grade methanol at a flow rate of 0.5 mL min -1 and concentrated to 1.0 mL volume, over a stream of nitrogen gas, in amber-colored auto-sampler vials for further analysis.

Liquid Chromatography/Mass Spectrometry

A triple quadrupole mass spectrometer (Micromass Quattro II, Waters Corp., Milford, Mass.) equipped with an electrospray source and coupled with a liquid chromatograph (1100 Series, Agilent Technologies, Santa Clara, Cal.) was used. An isocratic run was made with a ratio of 83:7:10 for HPLC-grade methanol, buffer of acetic acid glacial 0.05 M and ammonium hydroxide 0.05 M (pH 5), and Milli-Q water, respectively. Separation was achieved on an Agilent C-18 zorbax column, maintained at a temperature of 40°C. All aqueous solutions were filtered through a 0.22 µ m nylon membrane filter (Fisher Scientific, Fair Lawn, N.J.). The flow rate was maintained at 200 µ L min -1 . Injection volume was 10 µ L. Concentration was calculated using a five-point calibration curve. A 13 min post-time was allocated for re-equilibration of the column. Total run time was 28 min. Data were accumulated in scan mode, and analyses were carried out using MassLynx software (version 3.5, Waters Corp., Milford, Mass.). All MS parameters were optimized prior to analysis.

Statistical Analysis

The results were analyzed by applying the General Linear Model (GLM) procedure of least squares means adjustment for multiple comparisons (Tukey test) using SAS (version 9.2, SAS Institute, Inc., Cary, N.C.). The CORR procedure of SAS was used to develop a covariance matrix between five variables: pharmaceutical removal, temperature, dissolved oxygen (DO), oxidation reduction potential (ORP), and pH.

Results and Discussion

Tracer Study

In subsurface flow designs, the hydraulic conductivity of the selected soil substrate plays a decisive role in dictating the retention time of the system. In the case of treatment wetlands with limited area, it becomes imperative to select a medium that maximizes retention time while ensuring that the flow can be maintained on a sustained basis. Considering this requirement, sandy soils are appropriate substrates. The bromide tracer study (table 3) conducted to determine the retention time for the HSSF system showed some variation in retention times among the three replicates. The mean retention time was computed as 4.06 days.

Table 3. Bromide mass recovery and retention time in the HSSF treatment wetlands.

Constructed

Wetland

Mass Recovery

of Bromide (%)

Retention Time

(days)

CW 1

72.47

3.80

CW 2

64.18

4.12

CW 3

67.35

4.25

Antibiotic Removal in Wetlands

Removal of the evaluated pharmaceutical compounds, in the treatment wetlands, for the different concentration levels used is shown in figure 4. As the concentration level treatments were employed in an ascending order, and an equilibration time of 7 days with the subsequent concentration level was provided before sampling was initiated, it can be assumed that the residual effect of previous levels on the results was negligible. The removal efficiency of the wetland for level 2 concentration (50 µ g L -1 ) was similar to that of level 1 (10 µ g L -1 ). However, a decrease in removal efficiency was observed in level 3 (100 µ g L -1 ). Re-employment of level 3, in the following summer season, showed a marginal increase in the removal of the three antibiotics compared to the removal rates observed in the previous year. With further increase in concentration to levels 4 (500 µ g L -1 ) and 5 (1000 µ g L -1 ), a decreasing trend was observed.

The removal efficiencies observed at different levels did not differ significantly for most levels. Only levels 1 versus 5 (p < 0.01), levels 1 versus 6 (p < 0.05), and levels 2 versus 5 (p < 0.05) showed significant variation in their removal efficiencies. Level 3 (100 µ g L -1 ) was put to use in both years but at different times of the year. This provided an opportunity to directly observe temperature effects on the CWs' pollutant removal efficiency. The higher temperature regime showed greater removal rates but not significantly different from the lower temperature regime rates recorded in the previous year. In figure 4, only the higher temperature results are presented for level 3. In general, no clear trend was observed that would have explicitly reflected a pollutant concentration effect on the treatment potential of the tested wetland.

A significant difference existed between the removal efficiency of monensin and that of the other two antibiotics (p < 0.0001), but not between narasin and salinomycin (p > 0.05). Narasin showed the highest removal, and monensin showed the lowest. Among the ionophoric compounds evaluated, the lowest removal rate for monensin and the highest for narasin followed the sorption and degradation trends reported in earlier studies. Hussain and Prasher (2011) determined the respective value ranges for K d and log K oc of these compounds to be 38.24 to 8.52 L kg -1 and 4.91 to 4.26 L kg -1 , respectively, for sandy soil. These high values indicate that sorption could have been the dominant removal mechanism for this soil. According to laboratory findings with respect to the biodegradation potential of these antibiotics, the half-lives for monensin, salinomycin, and narasin ranged between 6.7 and 7.0 days (unpublished data). Considering the average retention time of 4.1 days, microbial dissipation might have also contributed to the overall removal.

The preferred microbial degradation route for pharmaceutical compounds is generally aerobic. Recent work has shown the relative effectiveness of aerobic environments in removing pharmaceuticals from wastewaters (Barceló et al., 2008). Therefore, it is possible that aerobic microbial degradation may have been active near the water-air interface and/or within the root zone. Although an anaerobic degradation pathway has not been determined as the preferred metabolic route for most pharmaceuticals (Kunkel and Radke, 2008; Musson et al., 2010), considering the largely anoxic environment of wetland systems, anaerobic degradation could also play a role in the overall removal.

Wetland studies conducted on HSSF systems have reported removal under anaerobic environments for certain pharmaceuticals (Ávila et al., 2010). Ibuprofen showed preference for aerobic biodegradation, while naproxen and diclofenac were efficiently removed (93%) under anaerobic conditions.

Microbial Factor

HSSF wetlands are expected to provide an enhanced pollutant-substrate contact opportunity, thereby augmenting the contribution of sorption and microbial degradation in overall removal. However, no significant statistical correlation of removal efficiency with environmental indicators like temperature, DO, ORP, or pH was observed in this study. The mean range of the detected values of these parameters is given in table 4. Nevertheless, the raw data trends for these parameters did imply a role for microbial degradation in the overall removal of ionophores. A gradual increase in pH during each season was observed at the outlet of the wetlands; this change could be associated with microbial activity, entailing that pH may be an indirect indicator of microbial activity. McKinley and Vestal (1984) stated that an increase in pH could be the result of microbial activity. Observed DO values for the three replicates were very close to zero, suggesting that what little oxygen was being furnished through the plant roots was being consumed very quickly by the microbial community (Camacho et al., 2007). It would be pertinent to mention that low values for DO should not be interpreted as unavailability of oxygen. An earlier study (Bezbaruah and Zhang, 2004) showed that a DO micro-profile can be present between root surfaces and bulk liquid. Lack of strong correlation of DO with temperature measurements, however, could be tied to the relatively low seasonal temperature 20.6°C (13°C to 26°C). Other studies have recorded higher DO values for HSSF wetlands (Coleman et al., 2001). Across both years, ORP values ranged between -9.57 and -12.13 mV, going into positive digits only during the initial time period of the study. This ORP trend indicates that these HSSF systems can be considered as facultatively anaerobic (Kadlec and Wallace, 2008).

Table 4. Average detected range of relevant soil parameters for the two-year study (standard errors are shown in parentheses).

Parameter

2008

2009

ORP (mV)

-14.57 ( ± 1.82)

-12.3 ( ± 2.35)

DO (mg L -1 )

0.27 ( ± 0.06)

0.38 ( ± 0.09)

Temp. (°C)

17.27 ( ± 1.31)

17.57 ( ± 1.17)

pH

7.32 ( ± 0.03)

7.18 ( ± 0.05)

Sorption and physical/biological degradation are considered to be the most common dissipation pathways for organic pollutants under a treatment wetland environment. In the current experiment, physical degradation was not considered an important removal process because (1) antibiotics are designed to resist hydrolysis (Naumova et al., 2010) and (2) due to the subsurface system employed in the tested CWs, photolysis was likely to be negligible. Moreover, our laboratory studies estimated a photodegradation half-life of 55.1, 40.1, and 37.2 days for monensin, salinomycin, and narasin, respectively (unpublished data). Considering the HSSF system and retention time of 4.1 days, appreciable influence of photodegradation was not likely. Based on these findings, sorption and biodegradation may be considered the dominant removal mechanisms in the evaluated CWs.

SW9213_files/image9.gif SW9213_files/image10.gif SW9213_files/image11.gif SW9213_files/image12.gif SW9213_files/image13.gif

Figure 4. Level-wise average weekly removal efficiency of the HSSF wetlands for the three antibiotics used: (a) first level (10 µ g L -1 ), (b) second level (50 µ g L -1 ), (c) third level (100 µ g L -1 ), (d) fourth level (500 µ g L -1 ), and (e) fifth level(1000 µ g L -1 ) levels, respectively. Error bars indicate standard errors from three replications.

Quantitative Removal by the Wetland

The quantitative removal achieved for each concentration level, with respect to the mass loadings of antibiotics, is given in figure 5. Overall mean removal efficiency ranges, calculated on quantitative terms across all levels, were 36.79% to 42.85%, 40.74% to 56.49%, and 41.88% to 57.08% for monensin, salinomycin, and narasin, respectively. Salinomycin showed a relatively higher removal efficiency than what was expected on the basis of laboratory-determined sorption and biodegradation potential. According to the laboratory findings, the removal by sorption or degradation of salinomycin should have been significantly lower than that of narasin (Hussain and Prasher, 2011; unpublished data). However, comparable removal results for salinomycin and narasin imply that it may be possible that some unevaluated dissipatory mechanism, such as anaerobic degradation, might be active for salinomycin in the wetland. The mass removal observed for different levels differed significantly for all levels; however, on percent removal basis, significant variation was only recorded for levels 1 versus 5 (p < 0.01), levels 1 versus 6 (p < 0.05), and levels 2 versus 5 (p < 0.05). Based on these findings, it can be asserted that, out of the three antibiotics, monensin is most likely to be detected in effluent waters. These results are in accordance with earlier reports of finding monensin in environmental waters (Lissemore et al., 2006; Watanabe et al., 2008). Other studies conducted on HSSF wetlands systems have detected variable removal rates for different compounds (Hijosa-Valsero et al., 2010a).

SW9213_files/image14.gif SW9213_files/image15.gif SW9213_files/image16.gif SW9213_files/image17.gif SW9213_files/image18.gif SW9213_files/image19.gif

Figure 5. Average total removal for each concentration level of the three antibiotics used: (a) first level (10 µ g L -1 ), (b) second level (50 µ g L -1 ), (c) third level (100 µ g L -1 ), (d) fourth level (500 µ g L -1 ) and (e) fifth level (1000 µ g L -1 ).

Conclusions

An investigation was carried out to estimate the removal efficiency of an HSSF treatment wetland, using a sandy soil substrate, for three ionophoric antibiotics: monensin, salinomycin, and narasin. Monensin was judged to be the most mobile, with an average removal efficiency of 39.82%, whereas salinomycin and narasin were not found to be statistically different in their removal potential in the CW. Quantitatively, narasin showed the highest removal efficiency (49.48%), closely followed by salinomycin (48.62%). The effect of the applied antibiotic concentration level on the wetland removal efficiency was found to be significant only for levels 1 (10 µ g L -1 ) versus 5 (500 µ g L -1 ) (p < 0.01), levels 1 versus 6 (1000 µ g L -1 ) (p < 0.05), and levels 2 (50 µ g L -1 ) versus 5 (p < 0.05). The effect of various environmental parameters including temperature, DO, ORP, and pH could not be significantly related to the removal rates, although raw data trends suggested a possible microbial contribution to overall antibiotics removal. The removal trends were in line with the findings of biodegradation and sorption laboratory studies carried out earlier.

REFERENCES

AHI. 2007. Trends in sales of life-saving animal medicines continue. Washington, D.C.: Animal Health Institute. Available at: http://ahi.timberlakepublishing.com/files/Media%20Center/Antibiotic%20use%202006.pdf.

APHA. 1992. Standard Methods for the Examination of Water and Wastewater . Washington, D.C.: American Public Health Association.

Ávila, C., A. Pedescoll, V. Matamoros, J. M. Bayona, and J. García. 2010. Capacity of a horizontal subsurface flow constructed wetland system for the removal of emerging pollutants: An injection experiment. Chemosphere 81(9): 1137-1142.

Barceló, D., M. Petrovic, V. Matamoros, and J. Bayona. 2008. Behavior of emerging pollutants in constructed wetlands. In Emerging Contaminants from Industrial and Municipal Waste , 199-217. Berlin, Germany: Springer.

Bezbaruah, A. N., and T. C. Zhang. 2004. pH, redox, and oxygen microprofiles in rhizosphere of bulrush ( Scirpus validus ) in a constructed wetland treating municipal wastewater. Biotech. and Bioeng . 88(1): 60-70.

Camacho, J. V., A. D. L. Martínez, R. G. Gómez, and J. M. Sanz. 2007. A comparatives study of five horizontal subsurface flow constructed wetlands using different plant species for domestic wastewater treatment. Environ. Tech. 28(12): 1333-1343.

Carlson, J. C., and S. A. Mabury. 2006. Dissipation kinetics and mobility of chlortetracycline, tylosin, and monensin in an agricultural soil in Northumberland County, Ontario, Canada . Environ. Toxicol. Chem. 25(1): 1-10.

Carmosini, N., and L. Lee. 2008. Sorption and degradation of selected pharmaceuticals in soil and manure. In Fate of Pharmaceuticals in the Environment and in Water Treatment Systems , 139-165. Boca Raton, Fla.: CRC Press.

Chapman, H. 2003. Removal of endocrine disruptors by tertiary treatments and constructed wetlands in subtropical Australia. Water Sci. and Tech. 47(9): 151-156.

Coleman, J., K. Hench, K. Garbutt, A. Sexstone, G. Bissonnette, and J. Skousen. 2001. Treatment of domestic wastewater by three plant species in constructed wetlands. Water, Air, and Soil Pollution 128(3-4): 283-295.

Conkle, J. L., J. R. White, and C. D. Metcalfe. 2008. Reduction of pharmaceutically active compounds by a lagoon wetland wastewater treatment system in southeast Louisiana. Chemosphere 73(11): 1741-1748.

Conkle, J. L., C. Lattao, J. R. White, and R. L. Cook. 2010. Competitive sorption and desorption behavior for three fluoroquinolone antibiotics in a wastewater treatment wetland soil. Chemosphere 80(11): 1353-1359.

Davis, J. G., C. C. Truman, S. C. Kim, J. C. Ascough II, and K. Carlson. 2006. Antibiotic transport via runoff and soil loss. J. Environ. Qual. 35(6): 2250-2260.

Dolliver, H., and S. Gupta. 2008a. Antibiotic losses from unprotected manure stockpiles. J. Environ. Qual. 37(3): 1238-1244.

Dolliver, H., and S. Gupta. 2008b. Antibiotic losses in leaching and surface runoff from manure-amended agricultural land. J. Environ. Qual. 37(3): 1227-1237.

EFSA. 2004a. Opinion of the scientific panel on additives and products or substances used in animal feed on the request of the commission on the reevaluation of coccidiostat elancoban in accordance with article 9G of Council Directive 70/524/EEC. EFSA Journal 42: 1-61.

EFSA. 2004b. Opinion of the scientific panel on additives and products or substances used in animal feed on a request from the commission on the safety and the efficacy of product "BIO-COX 120G" as feed additive in accordance with Council Directive 70/524/EEC. EFSA Journal 75: 1-51.

EFSA. 2004c. Opinion of the scientific panel on additives and products or substances used in animal feed on a request from the commission on the evaluation of coccidiostat Kokcisan 120G. EFSA Journal 63: 1-41.

EFSA. 2007. Opinion of the scientific panel on contaminants in the food chain on a request from the European Commission on cross-contamination of non-target feeding stuffs by narasin authorised for use as a feed additive. EFSA Journal 552: 1-35.

Elanco. 2004. Material safety data sheet: Maxiban Premix, chemical product and company. Greenfield, Ind.: Elanco, the Animal Health Division of Eli Lilly and Company.

Fisher, P. M. J., and R. Scott. 2008. Evaluating and controlling pharmaceutical emissions from dairy farms: A critical first step in developing a preventative management approach. J. Cleaner Prod. 16(14): 1437-1446.

Gray, J. L., and D. L. Sedlak. 2005. The fate of estrogenic hormones in an engineered treatment wetland with dense macrophytes. Water Environ. Res. 77(1): 24-31.

Hansen, M., K. A. Krogh, E. Björklund, B. Halling-Sřrensen, and A. Brandt. 2009. Environmental risk assessment of ionophores. TrAC Trends in Anal. Chem. 28(5): 534-542.

Hijosa-Valsero, M., V. Matamoros, J. Martín-Villacorta, E. Bécares, and J. M. Bayona. 2010a. Assessment of full-scale natural systems for the removal of PPCPs from wastewater in small communities. Water Res. 44(5): 1429-1439.

Hijosa-Valsero, M., V. Matamoros, R. Sidrach-Cardona, J. Martin-Villacorta, E. Bécares and J. M. Bayona. 2010b. Comprehensive assessment of the design configuration of constructed wetlands for the removal of pharmaceuticals and personal care products from urban wastewaters. Water Res. 44(12): 3669-3678.

Hussain, S., and S. Prasher. 2011. Understanding the sorption of ionophoric pharmaceuticals in a treatment wetland. Wetlands 31(3): 563-571.

Hussain, S., S. Prasher, M. Chenier, and G. Arya. 2011. Removal of nitrate-N by antibiotic exposed bacterial isolates from constructed wetlands. World J. Microbiol. Biotech. 27(9): 2061-2069.

Kadlec, R. H., and S. Wallace. 2008. Treatment Wetlands . 2nd ed. Boca Raton, Fla.: CRC Press.

Kim, S. C., and K. Carlson. 2006. Occurrence of ionophore antibiotics in water and sediments of a mixed-landscape watershed. Water Res. 40(13): 2549-2560.

Kim, S. H., M. Fan, S. O. Prasher, R. M. Patel, and S. A. Hussain. 2011. Fate and transport of atrazine in a sandy soil in the presence of antibiotics in poultry manures. Agric. Water Mgmt. 98(4): 653-660.

Kosolapov, D. B., P. Kuschk, M. B. Vainshtein, A. V. Vatsourina, A. Wiessner, M. Kastner, and R. A. Muller. 2004. Microbial processes of heavy metal removal from carbon-deficient effluents in constructed wetlands. Eng. Life Sci. 4(5): 403-411.

Kuchta, S. L., A. J. Cessna, J. A. Elliot, K. M. Peru, and J. V. Headley. 2009. Transport of lincomycin to surface and ground water from manure-amended cropland. J. Environ. Qual. 38(4): 1719-1727.

Kunkel, U., and M. Radke. 2008. Biodegradation of acidic pharmaceuticals in bed sediments: Insight from a laboratory experiment. Environ. Sci. Tech. 42(19): 7273-7279.

Lightbody, A. F., H. M. Nepf, and J. S. Bays. 2009. Modeling the hydraulic effect of transverse deep zones on the performance of short-circuiting constructed treatment wetlands. Ecol. Eng. 35(5): 754-768.

Lissemore, L., C. Hao, P. Yang, P. K. Sibley, S. Mabury, and K. R. Solomon. 2006. An exposure assessment for selected pharmaceuticals within a watershed in southern Ontario. Chemosphere 64(5): 717-729.

Matamoros, V., C. Arias, H. Brix, and J. M. Bayona. 2007. Removal of pharmaceuticals and personal care products (PPCPs) from urban wastewater in a pilot vertical flow constructed wetland and a sand filter. Environ. Sci. and Tech. 41(23): 8171-8177.

Matamoros, V., A. Caselles-Osorio, J. Garcia, and J. M. Bayona. 2008a. Behaviour of pharmaceutical products and biodegradation intermediates in horizontal subsurface flow constructed wetland. A microcosm experiment. Sci. Total Environ. 394(1): 171-176.

Matamoros, V., J. Garcia, and J. M. Bayona. 2008b. Organic micropollutant removal in a full-scale surface flow constructed wetland fed with secondary effluent. Water Res. 42(3): 653-660.

McKinley, V. L., and J. R. Vestal. 1984. Biokinetic analyses of adaptation and succession: Microbial activity in composting municipal sewage sludge. Appl. Environ. Microbiol. 47(5): 933.

Musson, S. E., P. Campo, T. Tolaymat, M. Suidan, and T. G. Townsend. 2010. Assessment of the anaerobic degradation of six active pharmaceutical ingredients. Sci. Total Environ. 408(9): 2068-2074.

Naumova, L., V. Batalova, G. Mokrousov, E. Didenko, and A. Solodkaya. 2010. Study of sorption and catalytic activity of peat-based composition material in relation to organic pollutants in waters. Russian J. Appl. Chem. 83(3): 395-399.

Onesios, K., J. Yu, and E. Bouwer. 2009. Biodegradation and removal of pharmaceuticals and personal care products in treatment systems: A review. Biodegradation 20(4): 441-466.

Park, N., B. J. Vanderford, S. A. Snyder, S. Sarp, S. D. Kim, and J. Cho. 2009. Effective controls of micropollutants included in wastewater effluent using constructed wetlands under anoxic condition. Ecol. Eng. 35(3): 418-423.

Ramaswamy, J., S. O. Prasher, R. M. Patel, S. A. Hussain, and S. F. Barrington. 2010. The effect of composting on the degradation of a veterinary pharmaceutical. Bioresource Tech. 101(7): 2294-2299.

Schlüsener, M. P., and K. Bester. 2006. Persistence of antibiotics such as macrolides, tiamulin, and salinomycin in soil. Environ. Pollution 143(3): 565-571.

Shilton, A. N., and J. N. Prasad. 1996. Tracer studies of a gravel bed wetland. Water Sci. Tech. 34(3-4): 421-425.

Sim, C. H., M. K. Yusoff, B. Shutes, S. C. Ho, and M. Mansor. 2008. Nutrient removal in a pilot and full scale constructed wetland, Putrajaya city, Malaysia. J. Environ. Mgmt. 88(2): 307-317.

Thiele-Bruhn, S. 2003. Pharmaceutical antibiotic compounds in soils: A review. J. Plant Nutrition Soil Sci . 166(2): 145-167.

Van de Moortel, A. M. K., D. P. L. Rousseau, F. M. G. Tack, and N. De Pauw. 2009. A comparative study of surface and subsurface flow constructed wetlands for treatment of combined sewer overflows: A greenhouse experiment. Ecol. Eng. 35(2): 175-183.

Watanabe, N., T. H. Harter, and B. A. Bergamaschi. 2008. Environmental occurrence and shallow groundwater detection of the antibiotic monensin from dairy farms. J. Environ. Qual. 37(5): 78-85.

Yates, C. R., and S. O. Prasher. 2009. Phosphorus reduction from agricultural runoff in a pilot-scale surface flow constructed wetland. Ecol. Eng. 35(12): 1693-1701.